Smoke test: validate demand without code before building

Table of Contents

• Smoke test methods, compared
• Why surveys mislead
• How to build the smoke test, step by step

  • 1. The fake door landing page
  • 2. Qualified traffic
  • 3. The preliminary CAC
• Two examples of how it would look
• 10-day plan

One of the most expensive mistakes when starting a business is building the full product before knowing whether anyone will pay for it. Surveys do not help much: almost everyone says "I would buy it" when they do not have to pull out a card.

A smoke test swaps the question for a behavior: you publish a real offer, with a price and a buy button, and measure how many people try to buy. With a landing page built without code and a small ad budget, you can have an answer in about ten days.

No ad budget? This guide uses 50 to 100 USD of paid traffic. If you want to do it with zero dollars, with interviews and organic traffic, follow the 14-day plan to validate a business idea without money.

Smoke test methods, compared

Method	What it measures	No-code tools	Rough signal	User friction
Fake door landing page	Click on "Buy" or "Start trial"	Framer, Webflow, Typeform	Several out of every 100 visitors click buy	Low
Refundable deposit pre-sale	Real payment of a token amount	Stripe, Lemon Squeezy, Mercado Pago	At least some visitors pay	High: removes false positives
Concierge or "Wizard of Oz"	Real use of a service done by hand behind the scenes	Airtable, Make, Zapier, WhatsApp Business	Users who come back week after week	Medium
B2B pre-sale	Letters of intent or signed pilots	DocuSign, a pitch deck	3 clients sign a pilot	Very high: requires meetings

Thresholds vary a lot depending on price, niche and traffic quality. Use them as a reference, not as a law.

Why surveys mislead

The book The Mom Test, by Rob Fitzpatrick, explains it well: if you ask "would you pay for this?", people answer what they think you want to hear. Between saying "I love it, I would buy it" and entering your card details there is a huge leap, and most ideas fall right there.

"Yes, I love it, I would buy it"  ◄── Surveys and interviews (false positive)
              │
              ▼
     [ THE PAYMENT LEAP ]           ◄── Your card or money has to come out
              │
              ▼
  Payment made / deposit            ◄── Payment intent validated

A well-run smoke test measures behavior on the other side of that leap.

Before launching it, lay out your hypothesis in the Business Model Canvas and define what progress your customer is seeking with the Jobs to be Done guide.

How to build the smoke test, step by step

1. The fake door landing page

The page has to look like a real product, ready to use.

• Headline centered on the problem. A headline about the main pain, a subtitle that explains the solution, and screenshots or mockups (Figma or Framer both work).
• Visible prices. Show concrete plans (for example, 19 and 49 USD per month). The price filters out the curious who are only looking for something free, and it makes the click mean something.
• The moment of truth. When the visitor clicks "Buy" or "Start trial", there are two paths:
  1. Waitlist: an honest message explaining that the product is about to launch and offering early access with a discount in exchange for an email.
  2. Deposit pre-sale: a simple checkout (Lemon Squeezy, Stripe or Mercado Pago) that charges a small, refundable deposit to reserve a spot, making clear it is a pre-sale. If you do not know the differences between these platforms, check the payment gateway comparison for LATAM.

2. Qualified traffic

Without enough traffic, the percentages mean nothing. For a first reading, aim for 300 to 500 visits from your niche, with 50 to 100 USD on Google Search or Meta.

• Google Search: words with buying intent ("billing software for dentists in Colombia"), not generic terms.
• Targeting: if the campaign does not convert, first check whether you are attracting your ideal customer profile or just curious people.

3. The preliminary CAC

Besides telling you whether the idea interests people, the smoke test gives you a first estimate of the cost of acquiring a customer (CAC):

Preliminary CAC = ad spend ÷ number of high-intent conversions (clicks on buy or deposits)

Illustrative example:

• Google Ads spend: 100 USD
• Qualified visits: 400 (about 0.25 USD per click)
• Clicks on "Buy 29 USD/month plan": 36 (9%)
• Preliminary CAC: 100 ÷ 36 ≈ 2.78 USD per person with purchase intent

Heads up: a click is not a customer. If only one in four of those 36 ended up paying, the real CAC would be about 11 USD. Even so, against a price of 29 USD per month, the numbers look healthy. Run the math with your costs in the profit margin calculator.

Two examples of how it would look

Debt collection micro-SaaS for property managers. Hypothesis: small property managers would pay 39 USD per month to automate payment reminders over WhatsApp. A landing page is built in Framer in an afternoon and 60 USD is invested in Meta targeting managers in two cities. If a meaningful share of visits tries to buy, the next step is a no-code MVP with Make and the WhatsApp API.

B2B technical service on subscription (concierge). Hypothesis: SMEs would pay a monthly subscription for preventive maintenance of their equipment. The offer is published with no platform behind it: when an order comes in, the founders coordinate an independent technician by hand and charge through a payment link. If clients renew month after month, there is evidence to build the platform or to go raise venture capital.

10-day plan

1. Day 1: offer and price. Define the value proposition and the launch price.
2. Days 2–3: landing page. Design it in Framer or Webflow, with the buy button connected to a waitlist or a refundable pre-sale.
3. Days 4–7: campaign. Invest 50 to 100 USD in Google Search or Meta to reach 300 or more visits from your niche.
4. Days 8–10: analysis and decision.
   • If a clear share of visitors tries to buy (as a reference, more than 5%), you have a green light for a no-code MVP or one built with an assistant like Cursor or Claude Code.
   • If almost nobody tries (as a reference, less than 2%), adjust the message, price or segment, or pivot before spending on development.
   • In any case, if you collected deposits and decide not to build, refund them.